enable creation of non-existing fields
[MojoFacets.git] / lib / MojoFacets / Data.pm
1 package MojoFacets::Data;
2
3 use strict;
4 use warnings;
5
6 use base 'Mojolicious::Controller';
7
8 use Data::Dump qw(dump);
9 use File::Slurp;
10 use JSON;
11 use Encode;
12 use locale;
13 use File::Find;
14 use Storable;
15
16 our $loaded;
17 our $filters;
18
19 sub index {
20         my $self = shift;
21
22         my $path = $self->app->home->rel_dir('data');
23         die "no data dir $path" unless -d $path;
24
25         my @files;
26         find( sub {
27                 my $file = $File::Find::name;
28                 if ( -f $file && $file =~ m/\.(js(on)?|txt)$/ ) {
29                         $file =~ s/$path\/*//;
30                         push @files, $file;
31                 } else {
32                         warn "IGNORE: $file\n";
33                 }
34         }, $path);
35
36         @files = sort { lc $a cmp lc $b } @files;
37         my $size;
38         $size->{$_} = -s "$path/$_" foreach @files;
39
40         my $dump_path;
41         $dump_path->{$_} = $self->_dump_path( $_ ) foreach @files;
42
43         $self->render(
44                 files => [ @files ],
45                 size => $size,
46                 loaded => $loaded,
47                 filters => $filters,
48                 dump_path => $dump_path,
49         );
50 }
51
52 sub _dump_path {
53         my ( $self, $name ) = @_;
54         my $dir = $self->app->home->rel_dir('data');
55         $name =~ s/^$dir//;
56         $name =~ s/\/+/_/g;
57         return '/tmp/mojo_facets.' . $name . '.storable';
58 }
59
60 sub _save {
61         my ( $self, $path ) = @_;
62
63         my $dump_path = $self->_dump_path( $path );
64         warn "save loaded to $dump_path";
65         my $info = $loaded->{$path};
66         store $info, $dump_path;
67
68         # sync timestamp
69         my $mtime = $loaded->{$path}->{mtime};
70         utime $mtime, $mtime, $dump_path;
71
72         warn $dump_path, ' ', -s $dump_path, " bytes\n";
73         return $dump_path;
74 }
75
76 sub _load_path {
77         my ( $self, $path ) = @_;
78
79         my $full_path = $self->app->home->rel_file( 'data/' . $path );
80         die "$full_path $!" unless -r $full_path;
81
82         my $dump_path = $self->_dump_path( $path );
83
84         if ( defined $loaded->{$path}->{data} ) {
85                 my $mtime = (stat($full_path))[9];
86                 return if $loaded->{$path}->{mtime} == $mtime;
87                 warn "reload $full_path, modified ", time() - $mtime, " seconds ago\n";
88         } elsif ( -e $dump_path ) {
89                 warn "dump_path $dump_path ", -s $dump_path, " bytes loading...\n";
90                 my $info = retrieve $dump_path;
91                 $loaded->{ $path } = $info;
92                 return;
93         }
94
95         # we could use Mojo::JSON here, but it's too slow
96 #       $data = from_json read_file $path;
97         my $data = read_file $full_path;
98         warn "# data snippet: ", substr($data,0,200);
99         my @header;
100         if ( $path =~ m/\.js/ ) {
101                 Encode::_utf8_on($data);
102                 $data = from_json $data;
103         } elsif ( $path =~ m/\.txt/ ) {
104                 my @lines = split(/\r?\n/, $data);
105                 $data = { items => [] };
106
107                 my $header_line = shift @lines;
108                 my $multiline = $header_line =~ s/\^//g;
109                 @header = split(/\|/, $header_line );
110                 warn "# header ", dump( @header );
111                 while ( my $line = shift @lines ) {
112                         $line =~ s/\^//g;
113                         chomp $line;
114                         my @v = split(/\|/, $line);
115                         while ( @lines && $#v < $#header ) {
116                                 $line = $lines[0];
117                                 $line =~ s/\^//g;
118                                 chomp $line;
119                                 my @more_v = split(/\|/, $line);
120                                 if ( $#v + $#more_v > $#header ) {
121                                         warn "short line: ",dump( @v );
122                                         last;
123                                 }
124                                 shift @lines;
125                                 $v[ $#v ] .= shift @more_v if @more_v;
126                                 push @v, @more_v if @more_v;
127
128                                 if ( $#v > $#header ) {
129                                         die "# splice $#header ", dump( @v );
130                                         @v = splice @v, 0, $#header;
131                                 }
132                         }
133                         my $item;
134                         foreach my $i ( 0 .. $#v ) {
135                                 $item->{ $header[$i] || "f_$i" } = [ $v[$i] ];
136                         }
137                         push @{ $data->{items} }, $item;
138                 }
139         } else {
140                 warn "file format unknown $path";
141         }
142
143         my $stats;
144
145         foreach my $e ( @{ $data->{items} } ) {
146                 foreach my $n ( keys %$e ) {
147                         $stats->{$n}->{count}++;
148                         my @v;
149                         if ( ref $e->{$n} eq 'ARRAY' ) {
150                                 $stats->{$n}->{array} += $#{ $e->{$n} } + 1;
151                                 @v = @{ $e->{$n} };
152                         } else {
153                                 @v = ( $e->{$n} );
154                         }
155
156                         foreach my $x ( @v ) {
157                                 $stats->{$n}->{numeric}++
158                                         if $x =~ m/^[-+]?([0-9]*\.[0-9]+|[0-9]+)$/;
159                                 $stats->{$n}->{empty}++
160                                         if length $x == 0; # faster than $x =~ m/^\s*$/;
161                         }
162
163                 }
164         }
165
166         foreach my $n ( keys %$stats ) {
167                 next unless defined $stats->{$n}->{array};
168                 delete $stats->{$n}->{array}
169                         if $stats->{$n}->{array} == $stats->{$n}->{count};
170         }
171
172         if ( ! @header ) {
173                 if ( defined $data->{header} ) {
174                         if ( ref $data->{header} eq 'ARRAY' ) {
175                                 @header = @{ $data->{header} };
176                         } else {
177                                 warn "header not array ", dump( $data->{header} );
178                         }
179                 }
180         }
181
182         @header =
183                 sort { $stats->{$b}->{count} <=> $stats->{$a}->{count} }
184                 grep { defined $stats->{$_}->{count} } keys %$stats
185                 unless @header;
186
187         warn dump($stats);
188
189         my $info = {
190                 header => [ @header ],
191                 stats  => $stats,
192                 full_path => $full_path,
193                 size => -s $full_path,
194                 mtime => (stat($full_path))[9],
195                 data => $data,
196         };
197
198         $loaded->{ $path } = $info;
199         $self->_save( $path );
200
201 }
202
203
204 sub load {
205         my $self = shift;
206
207         my @paths = $self->param('paths');
208         warn "# paths ", dump @paths;
209         $self->_load_path( $_ ) foreach @paths;
210
211         my $path = $self->param('path') || $self->redirect_to( '/data/index' );
212         warn "# path $path\n";
213         $self->session('path' => $path);
214         $self->_load_path( $path );
215
216         $self->session( 'header' => $loaded->{$path}->{header} );
217         if ( ! defined $loaded->{$path}->{columns} ) {
218                 $self->session( 'columns' => $loaded->{$path}->{header} );
219                 $self->session( 'order'   => $loaded->{$path}->{header}->[0] );
220                 $self->redirect_to( '/data/columns' );
221         } else {
222                 $self->session( 'columns' => $loaded->{$path}->{columns} );
223                 $self->session( 'order'   => $loaded->{$path}->{columns}->[0] );
224                 $self->redirect_to( '/data/items' );
225         }
226 }
227
228
229 sub _loaded {
230         my ( $self, $name ) = @_;
231         my $path = $self->session('path');
232         $self->redirect_to('/data/index') unless $path;
233         if ( ! defined $loaded->{$path}->{$name} ) {
234                 warn "$path $name doesn't exist in loaded ",dump( $loaded );
235                 $self->redirect_to('/data/index');
236         }
237         return $loaded->{$path}->{$name};
238 }
239
240
241 sub _checked {
242         my $self = shift;
243         my $checked;
244         $checked->{$_}++ foreach @_;
245 #       warn "# _checked ",dump($checked);
246         return $checked;
247 }
248
249
250 sub columns {
251     my $self = shift;
252
253         if ( $self->param('columns') ) {
254                 $self->_perm_array('columns');
255                 $self->redirect_to('/data/items');
256         }
257
258         my $stats = $self->_loaded( 'stats' ); # || $self->redirect_to( '/data/index' );
259
260         my @columns;
261         @columns = grep { defined $stats->{$_}->{count} } @{ $self->session('columns') } if $self->session('columns');
262
263         foreach my $c ( @{ $self->session( 'header' ) } ) {
264                 push @columns, $c unless grep { /^\Q$c\E$/ } @columns;
265         }
266
267     $self->render(
268                 message => 'Select columns to display',
269                 stats => $stats,
270                 columns => \@columns,
271                 checked => $self->_checked( $self->_perm_array('columns') ),
272         );
273 }
274
275 sub _perm_array {
276     my ($self,$name) = @_;
277
278         my @array = $self->param($name);
279         my $path  = $self->session('path');
280
281         if ( @array ) {
282                 $self->session($name => [ @array ]);
283                 $loaded->{$path}->{$name} = [ @array ];
284         } elsif ( my $session = $self->session($name) ) {
285                 if ( ref $session eq 'ARRAY' ) {
286                         @array = @$session;
287                 } else {
288                         die "$name not array ",dump($session);
289                 }
290         }
291         #warn "# $name ",dump @array;
292         return @array;
293 }
294
295 sub _perm_scalar {
296     my ($self,$name,$default) = @_;
297
298         my $scalar = $self->param($name);
299
300         if ( defined $scalar ) {
301                 $self->session($name => $scalar);
302         } else {
303                 $scalar = $self->session($name);
304         }
305
306         if ( ! defined $scalar ) {
307                 $scalar = $default;
308                 die "no default for $name" unless defined $scalar;
309                 $self->session($name => $scalar);
310         }
311
312         warn "# _perm_scalar $name ",dump $scalar;
313         return $scalar;
314 }
315
316 sub filter {
317         my $self = shift;
318
319         my $name = $self->param('filter_name') || die "name?";
320         my @vals = $self->param('filter_vals');
321
322         $self->_remove_filter( $name );
323         $self->_filter_on_data( $name, @vals ) if @vals;
324
325         $self->session( 'offset' => 0 );
326
327         $self->redirect_to('/data/items');
328 }
329
330 sub _filter_on_data {
331         my ( $self, $name, @vals ) = @_;
332
333         my $path = $self->session('path');
334
335         if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
336                 warn "filter $name not found in data set";
337                 return;
338         }
339
340         $filters->{$name} = [ @vals ];
341         warn "_filter_on_data $name ", $#vals + 1, " values on $path\n";
342
343         my $filter_hash;
344         $filter_hash->{$_}++ foreach @vals;
345
346         #warn "# filter_hash ",dump( $filter_hash );
347
348         my $items = $self->_loaded('data')->{items};
349
350         my $include_missing = defined $filter_hash->{_missing};
351         my $filtered_items;
352
353         foreach my $i ( 0 .. $#$items ) {
354
355                 if ( defined $items->[$i]->{$name} ) {
356                         foreach my $v ( @{ $items->[$i]->{$name} } ) {
357                                 if ( defined $filter_hash->{ $v } ) {
358                                         $filtered_items->{$i}++;
359                                 }
360                         }
361                 } elsif ( $include_missing ) {
362                         $filtered_items->{$i}++;
363                 }
364         }
365
366         #warn "# filter $name ",dump($filtered_items);
367
368         $loaded->{$path}->{filters}->{$name} = $filtered_items;
369         warn "filter $name with ", scalar keys %$filtered_items, " items created\n";
370 }
371
372 sub _data_items {
373         my ( $self, $all ) = @_;
374         my $data = $self->_loaded( 'data' );
375
376         return @{ $data->{items} } if $all == 1;
377
378         my $filters = $self->_current_filters;
379         my $filter_value;
380         foreach my $f ( keys %$filters ) {
381                 foreach my $n ( @{ $filters->{$f} } ) {
382                         $filter_value->{$f}->{$n} = 1;
383                 }
384         }
385         my @items = @{ $data->{items} };
386         @items = grep {
387                 my $i = $_;
388                 my $pass = 1;
389                 foreach my $n ( keys %$filter_value ) {
390                         if ( ! exists $i->{$n} ) {
391                                 if ( defined $filter_value->{$n}->{_missing} ) {
392                                         $pass = 1;
393                                         next;
394                                 } else {
395                                         $pass = 0;
396                                         last;
397                                 }
398                         }
399                         # and match any of values in element
400                         my $have_values = 0;
401                         foreach my $v ( @{ $i->{$n} } ) { # FIXME not array?
402                                 $have_values ||= 1 if defined $filter_value->{$n}->{$v};
403                         }
404                         if ( ! $have_values ) {
405                                 $pass = 0;
406                                 last;
407                         }
408                 }
409                 $pass;
410         } @items if $filter_value;
411         return @items;
412 }
413
414
415 sub _current_filters {
416         my $self = shift;
417         my $current_filters;
418         $current_filters->{ $_ } = $filters->{ $_ }
419                 foreach (
420                         grep { defined $filters->{ $_ } }
421                         @{ $self->_loaded('header') }
422                 );
423         #warn "# current_filters ",dump($current_filters);
424         return $current_filters;
425 }
426
427 sub _data_sorted_by {
428         my ( $self, $order ) = @_;
429
430         my $path = $self->session('path');
431
432         warn "_data_sorted_by $order from $path";
433
434         if ( defined $loaded->{$path}->{sorted}->{$order} ) {
435                 return $loaded->{$path}->{sorted}->{$order};
436         }
437
438         my $data = $self->_loaded( 'data' );
439         my $numeric = $self->_is_numeric($order);
440         my $missing = $numeric ? 0 : '';
441         no warnings qw(numeric);
442         my $nr = 0;
443         my @sorted = map {
444                 $_->[0]
445         } sort {
446                 if ( $numeric ) {
447                         $a->[1] <=> $b->[1]
448                 } else {
449                         $a->[1] cmp $b->[1]
450                 }
451         } map {
452                 [ $nr++, exists $_->{$order} ? join('', @{$_->{$order}}) : $missing ]
453         } @{ $data->{items} }
454         ;
455
456         warn "sorted: $order numeric: $numeric items: ", $#sorted + 1, "\n";
457         #warn "# sorted ",dump( @sorted );
458
459         $loaded->{$path}->{sorted}->{$order} = [ @sorted ];
460 }
461
462
463 sub items {
464         my $self = shift;
465
466         my $path = $self->session('path');
467         $self->redirect_to('/data/index') unless defined $loaded->{ $path };
468
469         my @columns = $self->_perm_array('columns');
470         $self->redirect_to('/data/columns') unless @columns;
471         my $order   = $self->_perm_scalar('order', $columns[0]);
472         my $sort    = $self->_perm_scalar('sort', 'a');
473         my $offset  = $self->_perm_scalar('offset', 0);
474         my $limit   = $self->_perm_scalar('limit', 20);
475         $self->_perm_scalar('show', 'table');
476
477         # fix offset when changing limit
478         $offset = int( $offset / $limit ) * $limit;
479
480         my $sorted = $self->_data_sorted_by( $order );
481
482         my @filter_names;
483         if ( $filters ) {
484                 foreach my $name ( keys %$filters ) {
485                         if ( ! defined $loaded->{$path}->{stats}->{ $name } ) {
486                                 warn "skip filter $name not found in $path\n";
487                                 next;
488                         }
489                         push @filter_names, $name;
490                 }
491                 warn "filter_names ",dump( @filter_names );
492                 foreach my $name ( @filter_names ) {
493                         next if ref $loaded->{$path}->{filters}->{$name} eq 'ARRAY';
494                         $self->_filter_on_data( $name, @{ $filters->{$name} } );
495                 }
496         }
497
498         my $all_filters = join(' ',sort @filter_names,'order:',$order);
499
500 #       warn "# all_filters $all_filters ", dump( $loaded->{$path}->{filtered}->{$all_filters} );
501
502         if ( ! defined $loaded->{$path}->{filtered}->{$all_filters} ) {
503
504                 my $path_filters = $loaded->{$path}->{filters};
505
506                 warn "create combined filter for $all_filters\n";
507
508                 my @filtered;
509                 foreach my $i ( 0 .. $#$sorted ) {
510                         my $pos = $sorted->[$i];
511
512                         if ( $#filter_names == -1 ) {
513                                 push @filtered, $pos;
514                                 next;
515                         }
516
517                         my $skip = 0;
518                         foreach ( @filter_names ) {
519                                 $skip ||= 1 if ! defined $path_filters->{$_}->{$pos};
520                         }
521                         next if $skip;
522
523                         push @filtered, $pos;
524                 }
525
526                 $loaded->{$path}->{filtered}->{$all_filters} = [ @filtered ];
527         }
528
529         my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
530                 if defined $loaded->{$path}->{filtered}->{$all_filters};
531
532         warn "all_filters $all_filters produced ", $#$filtered + 1, " items\n" if $filtered;
533
534         my $sorted_items;
535         my $data = $self->_loaded('data');
536         my $from_end = $sort eq 'd' ? $#$filtered : 0;
537         foreach ( 0 .. $limit ) {
538                 my $i = $_ + $offset;
539                 last unless defined $filtered->[$i];
540                 $i = $from_end - $i if $from_end;
541                 my $id = $filtered->[$i];
542                 push @$sorted_items,
543                 my $item = $data->{items}->[ $id ];
544                 $item->{_row_id} ||= $id;
545         }
546
547         warn "# sorted_items ", $#$sorted_items + 1, " offset $offset limit $limit order $sort";
548
549         $self->render(
550                 order => $order,
551                 offset => $offset,
552                 limit => $limit,
553                 sorted => $sorted_items,
554                 columns => [ @columns ],
555                 rows => $#$filtered + 1,
556                 numeric => { map { $_, $self->_is_numeric($_) } @columns },
557                 filters => $self->_current_filters,
558         );
559
560 }
561
562
563 sub order {
564         my $self = shift;
565         $self->session('order', $self->param('order'));
566         $self->session('sort', $self->param('sort'));
567         $self->redirect_to('/data/items');
568 }
569
570 sub _is_numeric {
571         my ( $self, $name ) = @_;
572
573         my $stats = $self->_loaded( 'stats' );
574
575         # sort facet numerically if more >50% elements are numeric
576         my $count = $stats->{$name}->{count};
577         $count   -= $stats->{$name}->{empty} if defined $stats->{$name}->{empty};
578         defined $stats->{$name}->{numeric} &&
579                 $stats->{$name}->{numeric} > $count / 2;
580 }
581
582 sub _remove_filter {
583         my ($self,$name) = @_;
584         warn "_remove_filter $name\n";
585
586         my $path = $self->session('path');
587
588         delete $filters->{$name};
589         delete $loaded->{$path}->{filters}->{$name};
590         warn "filters left: ", keys %{ $loaded->{$path}->{filters} };
591
592         foreach (
593                         grep { /\b$name\b/ }
594                         keys %{ $loaded->{$path}->{filtered} }
595         ) {
596                 delete $loaded->{$path}->{filtered}->{$_};
597                 warn "remove filtered cache $_";
598         }
599 }
600
601 sub facet {
602         my $self = shift;
603
604         my $path = $self->session('path') || $self->redirect_to( '/data/index' );
605
606         if ( my $name = $self->param('remove') ) {
607                 $self->_remove_filter( $name );
608                 $self->redirect_to( '/data/items' );
609         }
610
611         my $facet;
612         my $name = $self->param('name') || die "no name";
613
614         my $all = $self->_perm_scalar('all', 1);
615         my $data = $self->_loaded('data');
616
617         my $filters = $self->_current_filters;
618         my $all_filters = join(' ',sort keys %$filters,'order:',$self->session('order'));
619         my $filtered = $loaded->{$path}->{filtered}->{$all_filters}
620                 if defined $loaded->{$path}->{filtered}->{$all_filters};
621
622         if ( ! $filtered || $all ) {
623                 $filtered = [ 0 .. $#{ $data->{items} } ];
624                 warn "filter all values\n";
625         } else {
626                 warn "filter using $all_filters\n";
627         }
628
629         foreach my $i ( @$filtered ) {
630                 my $item = $data->{items}->[$i];
631                 if ( ! exists $item->{$name} ) {
632                         $facet->{ _missing }++;
633                 } elsif ( ref $item->{$name} eq 'ARRAY' ) {
634                         $facet->{$_}++ foreach @{ $item->{$name} };
635                 } else {
636                         $facet->{ $item->{$name} }++;
637                 }
638         }
639
640 #       warn "# facet $name ",dump $facet;
641
642         my $checked;
643         my @facet_names =
644                   $all                      ? keys %$facet
645                 : defined $filters->{$name} ? @{ $filters->{$name} }
646                 : keys %$facet;
647
648         $checked = $self->_checked( @{ $filters->{$name} } ) if defined $filters->{$name};
649
650         my $numeric = $self->_is_numeric($name);
651
652         my $sort = $self->param('sort');
653         # sort numeric facets with more than 5 values ascending
654         $sort ||= $numeric && $#facet_names > 4 ? 'a' : 'c';
655
656         @facet_names = sort {
657                 my $result;
658                 if ( $sort eq 'a' ) {
659                         $result = $numeric ? $a <=> $b : lc $a cmp lc $b;
660                 } elsif ( $sort eq 'd' ) {
661                         $result = $numeric ? $b <=> $a : lc $b cmp lc $a;
662                 } elsif ( $sort eq 'c' ) {
663                         $result = ( $facet->{$b} || -1 ) <=> ( $facet->{$a} || -1 )
664                 } else {
665                         warn "unknown sort: $sort";
666                         $result = $a cmp $b;
667                 }
668                 $result = $a cmp $b unless defined $result; # FIXME cludge for numeric facets with invalid data
669                 $result;
670         } @facet_names;
671
672         $self->render( name => $name, facet => $facet, checked => $checked,
673                 facet_names => \@facet_names, sort => $sort, numeric => $numeric,
674         );
675 }
676
677 sub edit {
678         my $self = shift;
679         my $content = $self->param('content');
680
681         my $i = $self->param('_row_id');
682         die "invalid _row_id ",dump($i) unless $i =~ m/^\d+$/;
683         my $path = $self->param('path') || die "no path";
684         my $name = $self->param('name') || die "no name";
685         my $status = 200; # 200 = OK, 201 = Created
686
687         if ( defined $loaded->{$path}->{data}->{items}->[$i] ) {
688                 $content =~ s/^\s+//s;
689                 $content =~ s/\s+$//s;
690                 my $v;
691                 if ( $content =~ /\xB6/ ) {     # para
692                         $v = [ split(/\s*\xB6\s*/, $content) ];
693                 } else {
694                         $v = [ $content ];
695                 }
696
697                 my $old = dump $loaded->{$path}->{data}->{items}->[$i]->{$name};
698                 my $new = dump $v;
699                 if ( $old ne $new ) {
700                         warn "# update $path $i $old -> $new\n";
701                         $loaded->{$path}->{data}->{items}->[$i]->{$name} = $v;
702
703                         if ( defined $loaded->{$path}->{sorted}->{$name} ) {
704                             delete $loaded->{$path}->{sorted}->{$name};
705                                 warn "# invalidate $path sorted $name\n";
706                         }
707
708                         foreach ( grep { m/$name/ } keys %{ $loaded->{$path}->{filtered} } ) {
709                             delete $loaded->{$path}->{filtered}->{$_};
710                                 warn "# invalidate $path filtered $_\n";
711                         }
712
713                         $status = 201; # created
714                         $self->session('save_path' => $path);
715
716                 } else {
717                         warn "# unchanged $path $i $old\n";
718                         $status = 304;
719                 }
720         } else {
721                 $content = "$path $i $name doesn't exist\n";
722                 $status = 404;
723         }
724
725         warn "# edit $status $content";
726
727         $self->render(
728                 status => $status,
729                 content => $content,
730         );
731 }
732
733
734 sub save {
735         my $self = shift;
736         my $path = $self->param('path');
737         $path  ||= $self->session('path');
738
739         my $dump_path = $self->_save( $path );
740         $self->session('save_path' => 0);
741
742         $self->redirect_to( '/data/items' );
743 }
744
745 1;